Your monthly close looks perfect on the surface. MRR beat the forecast by 15% and ARR is climbing steadily. You send the board deck out with confidence. Then you dig into the underlying trends. You realize that revenue surge isn't coming from healthy new acquisition. It is almost entirely driven by aggressive upsells to a few legacy accounts. Meanwhile, the large large group of enterprise customers you signed last quarter is already downgrading, which quietly spikes your churn rate.
Aggregated metrics give you the headline. Cohort analysis tells you the real story. When you track revenue using time based cohorts, you spot patterns big-picture numbers hide. You see exactly if your March enterprise customers are growing or if the newest cohort retains better than the last. This detailed view helps you forecast accurately and allocate resources where they actually drive value.
Why traditional revenue metrics miss trends
MRR and ARR help, but you only see what you have now. You don’t see where it came from or how stable it is.
Once expansion revenue enters the picture, things get tricky. NRR at 120% looks great. Dig a little and maybe 70% of your base revenue leaves each year. A handful of power users keep you afloat with upsells. Take our retention advice: “If you have 70% GRR and 100% NRR, you’re masking churn with upsells.”
This is where gross dollar retention (GDR) comes in. GDR strips out expansion and shows what percentage of your starting revenue you actually keep. Upsells can hide a high churn rate. GDR brings real results to the surface.
Still, even GDR and NRR can lead you astray if you only look at them in aggregate. Maybe your company-wide retention is 95%. That sounds healthy. But your newest cohorts could be exiting at 20% while your oldest customers are steady. Or maybe the legacy group is shrinking and new signups are sticking.
You need to see retention by vintage. Group customers by start date. Track how their revenue changes month by month. Cohort-based retention views do exactly that, and finance teams should always keep them in their models.
What is cohort analysis for finance teams
Cohort analysis for finance is simple. Group customers by start date. Track their revenue over time. Use contract start date, first invoice, or whatever fits your business.
This differs from the behavioral cohorts product teams use to check feature adoption or activation. For finance, you ask: “How much revenue did January 2026 signups bring in during month one? Month three? Month twelve?” Compare that to April cohort, May, and so on.
It creates a matrix:
- One axis: the cohort’s start month
- Other axis: time since each cohort began (“cohort age”)
- Each cell: revenue, retention rate, or seat count for the cohort at that point
This setup makes trends easy to spot across customer tier, region, or contract type. You can see if enterprise customers retain better than SMB. You can see if Q1 groups perform differently than Q4. You can set up your forecasts on a relative timeline (three months after sign-up, six months after sign-up) then roll those assumptions across every cohort moving forward.
Cohort modeling helps finance teams track and forecast how groups behave over time, whether it’s seat growth, usage, retention, revenue, or churn. Instead of a single MRR or ARR number, answer questions like “What’s the cumulative revenue from our March enterprise cohort?” and “Where’s growth accelerating or leveling out across segments?”. We cover more in our cohort modeling guide.
The key difference from standard retention dashboards is simple. Cohort tables are a planning tool. You build a model, plug it right into your financial plan, and update it as new data arrives.
Why finance teams should own this analysis
Cohort insights give finance sharper forecasts and smarter planning. It’s also the most accurate way to understand customer lifetime value.
Forecasts improve when you model by cohort vintage. If your Q1 2026 group kept 92% of revenue after six months and Q2 looks similar at month three, use those real signals. You can forecast Q2’s six-month retention using actual behavior from comparable groups rather than guessing.
Cohort analysis shows exactly where growth originates. Breaking revenue down by group reveals the mechanics behind your numbers:
- New customer acquisitions versus expansion of your current base
- Growth driven by upsells to fewer accounts
- New groups that are smaller but retain better
Each scenario changes how you plan.
Cohort views also help CFOs and FP&A leads test model assumptions effectively. Speculating 95% gross retention when your last three cohorts trend at 88% creates a gap you need to fix. Cohort data lets you update those numbers with facts instead of guesses.
Net revenue retention acts as the heartbeat of your financial model… Great finance teams use NRR to forecast cash flow without guesswork. Layering in cohort views gives you even more control. Dollar-based NRR tracks a specific group’s revenue over time, shows trends by group, and keeps the data clear.
Investors also care about these trends. NRR sets the pulse for your forecasts, planning, and business health updates. Investors and acquirers see specific retention data as a vital component of customer lifetime value. When you show retention by cohort and prove that new groups perform as well or better than before, you tell a stronger, more confident story than with a single company-wide stat.
How to build a cohort analysis table in Runway: step by step
Here’s how to set up a cohort retention model in Runway that tracks revenue by customer group each month.
- Step 1: Import and structure your data
Start by bringing in raw subscription or contract data from your CRM or billing. Each row should be a customer or contract, with start date, monthly revenue, customer tier, region, and any other detail you track. Import it into Runway’s Source data database. Set the time dimension to use the contract start month for modeling. - Step 2: Create a Cohorts database
This is where you move from transactions to groups. Create a Cohorts database in Runway. Set its source to your imported data. This gives you a database where each segment is a cohort, like “enterprise customers who signed up in March 2026.” - Bring in the cohort start month as a driver for time-based modeling.
- Include drivers to roll up: seats, contract value, revenue.
- Segment by cohort start (like contract start month). Add tier, geography, or product line if you want to split groups further.
- Step 3: Add cohort age and metrics
For each cohort, add a driver for age, such as months since it started. This applies assumptions to each month based on cohort age. Add drivers for churn, expansion revenue, and retained revenue. Roll them up from your data or set them as placeholders for now. - Step 4: Create an assumptions database for retention
Make the table forecastable. Set up an assumptions database. Segment by cohort age and by tier or segment. Add rows for each month of a cohort’s life (usually 12 or 24). Include a column for expected retention or growth at that age. That’s where you note: “At month 3, enterprise cohorts keep 95% of revenue. At month 6, they keep 92%,” and so on. - Step 5: Connect assumptions and cohorts for forecasting
Link the Cohorts and Assumptions databases. Each cohort uses the assumption row for its age and tier. Runway’s filters apply the right values month by month. Use different logic and multiple metrics as needed. This supports flexible, detailed forecasting that works for your model’s needs. - Step 6: Roll up to a retention overview
Create a high-level summary by rolling up all cohorts. Make a new database with its source set to your Cohort Model. Segment by tier or another key detail. Pull in driver metrics like revenue or seats. Now you have one row per segment for a clean performance overview, with easy drill-downs to specific assumptions, cohorts, or customers.
Runway’s modeling layer is built for this kind of analysis. Slice your model by department, customer type, or location. Write human-readable formulas for team, product, or region. Drill in instantly to see what makes up any number. Move from the big picture to detailed data all in one place.
How to read cohort patterns
Once your cohort retention table is set up, read it closely. Look for these three main patterns and note what they mean for your financial plan.
- Flattening curves
Cohort revenue starts at 100% in month zero, drops to 95% in month three, and then holds steady around 90-92%. This is good news. You lose some customers early (that’s normal), and those who stay are solid. Plan for a quick early drop, then steady rates after that. This is the result you want. - Steady decline
If retention falls every month (100%, 95%, 88%, 80%, 72%) your churn rate is a real risk. Customers don’t see lasting value, and revenue shrinks. Plan for continued decline until things stabilize. Focus on improving the product or customer experience before it affects new cohorts. Use GDR to spot this type of erosion early. Do this even when NRR looks fine because of expansion revenue. - Upward slopes
If a cohort’s revenue grows over time (100%, 105%, 112%, 120%) that’s strong expansion. Customers buy more seats, upgrade, or add products. This is positive, but check the mix. Is growth coming from a few big users or spread out? If it's concentrated, forecasts are riskier. Model expansion separately from retention. Track GRR and NRR as two clear drivers. This shows base stability and highlights where the upside is coming from.
Always compare cohorts. If your Q1 2026 group keeps more revenue at month six than Q4 2025 did, something improved. Maybe onboarding, pricing, or customer fit. If new cohorts retain less, dig in. Either you’re attracting the wrong customers, or something’s changed in your product or service.
With clear pattern spotting, you can make changes in your plan fast. Adjust retention assumptions, redirect budget to customer success, or shift growth targets.
Turn cohort analysis into ongoing forecasting in Runway
Cohort analysis isn’t a one-off report for the board. It’s a living input for your forecast. It updates continuously as new data comes in.
In Runway, when you set your Last Close date, the platform switches from imported actuals to forecast formulas. You see historical performance and future projections in one place. When you close January, January’s cohorts use actuals, and your model moves forecast logic out to February and beyond.
This continuous cycle powers strong planning. You’re not just reviewing the past. You’re using cohort data for every forecast update, budget review, and scenario plan.
Runway’s reporting layer turns cohort tables into dashboards and charts that update in real time. Share cohort retention views with your team, and they stay current as the model does. Build scenario comparisons like “What if the newest cohort retains 5 points higher than last time?” and see the financial impact across revenue, cash, and headcount.
This is how modern finance teams run. They don’t put cohort analysis on the side. They make it core to financial modeling and use it to drive fast, smart decisions.
If you’d like to see this in action, talk to us. We'll show you how to build a cohort retention model in Runway, connect it to your data, and make it part of your finance team’s day-to-day planning process.